A nueral netwoilc model for environmental predicadon: Case study for China

In this paper, the popular training method Back-propagation (BP) of Nueral Networks (NN) is modified by introducing two parameters into the procedure in order to speed learning and improve the convergence. The modified algorithm is named as BPGM and is successfully used in forecasting the future pollution generation and the investment demand for the environment protection in China. Based on the predication of NN model, the future situation of environment of China is analyzed and the investment in environment protection is compared with the average level of the world. Some policies suggestions are proposed.